Fourier Integral Estimate of the Failure Rate Function and Its Mean Square Error Properties,
Abstract
In this paper we introduce a new class of estimators of the failure rate function which are based on its Fourier transform. We show that these estimators are in fact kernel estimators based on the sinc function. They have, for a certain class of the failure rate functions a faster rate of convergence of the mean square error than those estimators based on other kernels. We attempt to explain the reason for this fast rate of convergence by pointing out the connection between a sinc kernel estimator and the jackknifing of kernel estimators. We make some concluding remarks on the meaning and the value of the results given in this paper. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 14, 1980
- Accession Number
- ADA099432
Entities
People
- Man-yuen Wong
- Nozer Singpurwalla
Organizations
- George Washington University